The present invention relates to a method and apparatus for estimating a fatigue level of an object to be estimated, and a program applicable to such method and apparatus.
Conventionally, this type of fatigue-level estimation apparatus has been disclosed by, for example, Japanese Patent Laid-open (KOKAI) No. 2002-65650. In this fatigue-level estimation apparatus, a heart rate and variations in heartbeat intervals in both of a non-driving state and a driving state are computed on the basis of a driver's heart rate signal. Both of the heart rate and the changes in the heartbeat intervals are developed onto a two-dimensional coordinate so as to form a map thereon so that a temporal transition trend of the map is observed. By considering this temporal transition trend, in addition to consideration about whether or not there is an instantaneous rise in the heart rate, a driver's fatigue level is estimated.
In the estimation process according to this conventional fatigue-level estimation apparatus, it is estimated that the mutually closer both of the maps in between the non-driving and driving states on the two-dimensional coordinate (of which two dimensions are the heart rate and the changes in the heartbeat intervals are), the higher the driver's fatigue level is.
However, the conventional fatigue-level estimation apparatus is configured such that the fatigue level is obtained through the comparison with past data. Thus, if the past data is unreliable, the reliability of resultant estimation of the fatigue level is forced to decrease. Since there are great differences among individuals in estimating the fatigue level, the estimation should be done every driver.
In addition, the conventional fatigue-level estimation apparatus has no configuration to take it consideration the factors, such as drivers' personal differences, drivers' physical conditions, and vehicle running conditions on roads. Thus, it is difficult for the conventional apparatus to determine whether a driver is really tired or not. Further, in order to avoid erroneous detection of the fatigue level, detection sensitivity can be lowered. But, if such a measure is taken, there is a possibility that failure in the detection is caused.